IBM has introduced new security and identity management capabilities designed to help businesses monitor, control and protect artificial intelligence agents operating across enterprise systems.
Announced at IBM TechXchange 2026 on 27 October, the developments bring together agent identity verification, credential management and governance tools to address the security challenges associated with increasingly autonomous AI technologies.
The new capabilities aim to give organisations greater visibility over AI agents, including who owns them, which applications they can access and what actions they are authorised to perform.
Strengthening security for autonomous AI
As businesses deploy AI agents to automate tasks and interact with sensitive information, managing their identities and permissions has become an increasingly important consideration.
IBM has introduced Agent Identity, which works alongside HashiCorp Vault to establish verified identities for AI agents and maintain records of their ownership, permissions and operational status.
The integration enables organisations to issue short-lived credentials, restrict access to approved systems and revoke permissions when an agent attempts an unauthorised action.
HashiCorp Vault 2.2 also introduces additional credential management capabilities, including automated credential renewal and support for post-quantum encryption.
Improving AI agent visibility and governance
IBM is also expanding its watsonx Orchestrate platform to help enterprises discover, evaluate and manage AI agents operating across different technology environments.
New capabilities are designed to identify agents built using third-party platforms, including Amazon Bedrock, Microsoft Foundry and Google Gemini.
The developments include isolated execution environments and controls that can shut down individual agents when required.
IBM has also introduced evaluation capabilities to help businesses assess agent performance, accuracy, compliance and operational costs.
Dinesh Nirmal, Senior Vice President of IBM Software, said that enterprises need to integrate AI agents into their existing operating models with appropriate security and accountability.
Supporting enterprise AI adoption
The announcement forms part of IBM’s wider strategy to provide businesses with technologies for securely deploying AI across applications, infrastructure and cloud environments.
Several capabilities, including Agent Identity, HashiCorp Vault 2.2 and watsonx Orchestrate evaluation tools, are generally available.
Other features, including agent discovery and containment within watsonx Orchestrate, remain in technical preview.
The developments highlight the growing importance of identity management, access controls and operational oversight as enterprises expand their use of autonomous AI systems.
